Bacterial resistance to microbicides - Applied and Environmental ...

AEM Accepted Manuscript Posted Online 30 January 2015 Appl. Environ. Microbiol. doi:10.1128/AEM.03843-14 Copyright © 2015, American Society for Microbiology. All Rights Reserved.

Applied and Environmental Microbiology 1

Bacterial resistance to microbicides: Development of a predictive protocol

2

Laura Knapp1, Alejandro Amézquita2, Peter McClure2, Sara Stewart2 and

3

Jean-Yves Maillard1*

4 5 6

1

Cardiff School of Pharmacy & Pharmaceutical Science, Cardiff, Wales, UK

2

Unilever Safety & Environmental Assurance Centre, Colworth Science Park,

Bedford, UK

7 8

Corresponding author:

9

Jean-Yves Maillard

10

Tel: (+44) 02920254828

11

Address: Cardiff School of Pharmacy and Pharmaceutical Sciences, Cardiff

12

University, Redwood Building, King Edward VII Avenue, Cardiff CF10 3NB, UK

13

Email: [email protected]

1


Abstract

15

Regulations dealing with microbicides in Europe and the United States are evolving and now

16

require data on the risk of resistance development in organisms targeted by microbicidal

17

products. There is no standard protocol to assess the risk of resistance development to

18

microbicidal formulations. This study aimed to validate the use of changes in microbicide

19

and antibiotic susceptibility as initial markers for predicting microbicide resistance and

20

cross-resistance to antibiotics. Three industrial isolates (Pseudomonas aeruginosa,

21

Burkholderia cepacia, Klebsiella pneumoniae) and two Salmonella enterica serovar

22

Typhimurium strains (SL1344 and 14028S) were exposed to a shampoo, a mouthwash, eye

23

make-up remover and the microbicides contained within these formulations (chlorhexidine

24

digluconate; CHG and benzalkonium chloride; BZC), under realistic, in-use conditions.

25

Baseline and post- exposure data were compared. No significant increases in minimum

26

inhibitory concentration (MIC) or minimum bactericidal concentration (MBC) were

27

observed in any strain after exposure to the three formulations. Increases in the MIC and

28

MBC of CHG and BZC of up to 100-fold were observed in SL1344 and 14028S but were

29

unstable. Changes in antibiotic susceptibility were not clinically significant.

30

The use of MICs and MBCs combined with antibiotic susceptibility profiling and stability

31

testing generated reproducible data that allowed for an initial prediction of microbicide

32

resistance development. These approaches measure characteristics that are directly relevant

33

to the concern over resistance and cross-resistance development following use of

34

microbicides. These techniques are low cost and high-throughput, allowing manufacturers to

35

provide data to support early assessment of risk of resistance development to regulatory

36

bodies promptly and efficiently.

37 38

Keywords: microbicides, resistance, predictive protocol, regulation

39

INTRODUCTION

2


Microbicides have been extensively used in the control of bacteria for decades, and

41

are commonly incorporated into a variety of products including disinfectants,

42

cosmetics, preservatives, pesticides and antiseptics. Despite this ever-increasing use,

43

bacteria generally remain susceptible to microbicides when they are used correctly.

44

However, the indiscriminate use of microbicides in a wide range of environments

45

has raised concerns about the selection of microbicide and antibiotic-resistant

46

bacteria (1, 2). Despite the establishment of the European Union (EU) biocidal

47

product regulation (BPR) (http://eur-

48

lex.europa.eu/LexUriServ/LexUriServ.do?uri=OJ:L:2012:167:0001:0123:EN:PDF

49

accessed 24th November 2014) to regulate the authorisation and use of biocidal

50

products throughout the EU, the total amount of microbicide use in the EU remains

51

unknown (2).

52 53

Of particular concern are formulations that contain microbicides at low

54

concentrations which may increase the risk of selection for resistance amongst target

55

or non-target microorganisms (2). Resistance or reduced susceptibility to

56

microbicides and/or antibiotics as a result of exposure to low microbicide

57

concentrations has been demonstrated extensively in the laboratory setting (3-7).

58

Despite the lack of in vivo or in situ studies reporting a link between microbicide

59

exposure and antibiotic resistance development, in vitro studies have clearly

60

demonstrated the possibility of microbicide and antibiotic resistance development in

61

bacteria. This has lead committees such as the Scientific Committee on Emerging

62

and Newly Identified Health Risks (SCENIHR) to produce reports and opinions on

63

the knowledge gaps and research concerns associated with resistance. In their 2010

64

opinion paper SCENIHR stated that data on microbicide usage are lacking together

3


with an understanding of the microbicides most at risk for the development of

66

bacterial resistance

67

(http://ec.europa.eu/health/scientific_committees/emerging/docs/scenihr_o_028.pdf,

68

accessed 24th November 2014). SCENIHR recommended the standardisation of

69

methodologies used to monitor resistance levels and suggested the development of a

70

standard protocol that could determine the risk of resistance development in a

71

particular microorganism as a result of microbicide exposure.

72 73

In support of the requirement for such a protocol, the new BPR (EU 528/2012) states

74

that it is a requirement of biocidal product manufacturers to provide information on

75

the likelihood of resistance development to their product in target organisms. In

76

particular the following articles state:

77

“(13) Active substances can, on basis of their intrinsic hazardous properties, be

78

designated as candidates for substitution with other active substances, whenever such

79

substances considered as efficient towards the targeted harmful organisms become

80

available in sufficient variety to avoid the development of resistances amongst

81

harmful organisms…”

82

“(25) … The use of low-risk biocidal products should not lead to a high risk of

83

developing resistance in target organisms.”

84

“(33) When biocidal products are being authorized, it is necessary to ensure that,

85

when properly used for the purpose intended, they are sufficiently effective and have

86

no unacceptable effect on the target organisms such as resistance...”.

87

In addition, the U.S. Food and Drug Administration (FDA) has also issued a

88

proposed rule to require manufacturers of antibacterial hand soaps and body washes

89

to demonstrate that their products are safe for long-term daily use, more effective

4


than plain soap and water in preventing the spread of certain infections and do not

91

select for resistance (http://www.gpo.gov/fdsys/pkg/FR-2013-12-17/pdf/2013-

92

29814.pdf accessed 24th November 2014). A standard protocol that could determine

93

the risk of resistance development would allow microbicidal product manufacturers

94

to provide this information to the BPR and FDA promptly and efficiently.

95

Our work focuses on the development of such a protocol and has involved the

96

assessment of several laboratory techniques that can be used to measure microbicide

97

resistance (e.g. minimum inhibitory concentration (MIC)/minimum bactericidal

98

concentration (MBC) determination, antibiotic susceptibility testing, and phenotype

99

stability testing) in terms of ease of use, high throughput, cost and reproducibility.

100

Our recommended protocol encompasses MIC, MBC and antibiotic susceptibility

101

determination combined with bacterial phenotype stability testing as initial markers

102

of bacterial microbicide resistance or antibiotic cross-resistance. This study aims to

103

validate the use of these techniques in a combination protocol with the testing of

104

three commercially available formulations and the corresponding active microbicides

105

contained therein.

106 107 108 109

MATERIALS AND METHODS

110

Bacterial strains. A range of Gram-negative bacteria was selected for the testing of

111

three antimicrobial formulations and the corresponding microbicides contained

112

within each formulation. The bacteria included Salmonella enterica serovar

113

Typhimurium strains SL1344 and 14028S (obtained from the University of

114

Birmingham, UK), Burkholderia cepacia (UL2P; Unilever culture collection, UK),

5


Klebsiella pneumoniae (UL13; Unilever culture collection, UK) and Pseudomonas

116

aeruginosa (UL-7P; Unilever culture collection, UK). The 3 Unilever strains were

117

selected as challenge organisms due to their routine use, propagation and handling in

118

Unilever laboratories.

119 120

Culture and storage of bacteria. Liquid cultures of all strains were grown in

121

tryptone soya broth (TSB) (Oxoid, Basingstoke, UK) at 37°C (± 1 °C). Strains were

122

stored on protect beads (Fisher Scientific, Loughborough, UK) at -80 °C (± 1 °C)

123

and restricted to a maximum of 2 subcultures from the original freezer stock prior to

124

exposure to a given microbicide. Test inocula were prepared from harvesting an

125

overnight TSB culture centrifuged at 5000 g for 10 min and re-suspended in

126

deionised water (diH20).

127 128

Formulations, actives and neutraliser. A mouthwash (2 mg/mL chlorhexidine

129

digluconate; CHG), eye make-up remover (1 mg/mL CHG) and a shampoo (5

130

mg/mL benzalkonium chloride; BZC) were tested. Selection of these products was

131

based on the fact that they are commonly used home and personal care products. The

132

microbicides CHG and BZC (Sigma-Aldrich, Dorset, UK), the only microbicides

133

contained within the three formulations, were also tested. The neutraliser used was

134

composed of Tween 80 (30 g/L) and Asolectin (3 g/L) (both Sigma-Aldrich, Dorset,

135

UK). Neutraliser efficacy for mouthwash, shampoo and eye make-up remover, and

136

toxicity towards all strains was determined as described previously (3).

137 138

Antimicrobial susceptibility testing

6


Suspension testing: Test strains were exposed to each formulation and each

140

microbicide at a concentration resulting in a 1-3 log10 reduction in CFU/mL, leaving

141

sufficient survivors for further antimicrobial susceptibility testing. Suspension tests

142

were carried out following the British Standard EN 1276 2009 protocol (8). Briefly,

143

bacterial suspensions in deionised water (diH20) produced from overnight cultures

144

were standardised to 1 x 108 CFU/mL. Suspensions were used within 15 minutes of

145

preparation. One mL of standardised suspension was added to 9 mL of the desired

146

formulation or active (diluted in diH20) at 1.25 times the required concentration.

147

Concentrations tested were as follows: 0.000125 mg/mL mouthwash/CHG, 0.00015

148

mg/mL shampoo/BZC and 1 mg/mL eye make-up remover/CHG. After exposure for

149

1 min (the estimated length of time spent using each formulation by the consumer), 1

150

mL of this suspension was removed and added to 9 mL of neutraliser. After

151

neutralisation, suspensions were centrifuged at 5000 g for 10 min and the

152

supernatant discarded. The remaining cells were then used in further antimicrobial

153

susceptibility testing experiments. S. enterica strains SL1344 and 14028S were also

154

exposed to low BZC and CHG concentrations ranging from 0.0001– 0.004 mg/mL

155

for 5 min.

156 157

Determination of the minimum inhibitory concentration (MIC). The MIC of each

158

formulation/microbicide was determined for all strains before and after suspension

159

test exposure to a given formulation/active, following the BS EN ISO: 20776-1 (9)

160

protocol. Briefly, a 96 well microtitre plate (Sterilin Ltd, Newport, UK) containing

161

doubling dilutions of a given formulation/active in TSB was set up. Concentration

162

ranges were as follows: Mouthwash/CHG 2 – 0.001 mg/mL, shampoo/BZC 1.25–

163

0.001 mg/mL, eye make-up remover/CHG 0.5 – 0.00048 mg/mL, CHG/BZC

7


(Salmonella strains only) 40 – 0.019 mg/mL. An overnight broth culture of each

165

strain was standardised to 1 x 108 CFU/mL and 50 µL volumes of this were added to

166

the microtitre plate. The plate was incubated for 24 h at 37°C. The MIC was defined

167

as the lowest concentration of a formulation/microbicide at which no bacterial

168

growth was observed visually on the microtitre plate. (Approximate cost to test one

169

microbicide and one bacterium in triplicate: < 1€).

170 171

Determination of the minimum bactericidal concentration (MBC). The MBC of

172

each formulation/microbicide was also determined before and after suspension test

173

exposure of each strain to a given formulation/active. Twenty µL of suspension was

174

removed from each well of the MIC microtitre plate where no bacterial growth was

175

observed and the two lowest formulation/active concentrations at which growth was

176

observed, and added to 180 µL of neutraliser. Twenty-five µl of this suspension was

177

then spotted on to tryptone soya agar (TSA) and incubated at 37°C for 24 h. The

178

minimum bactericidal concentration was defined as the lowest formulation/active

179

concentration where no bacterial growth was observed on the agar plate.

180

(Approximate cost to test one microbicide and one bacterium in triplicate: < 1 €).

181 182 183

Antibiotic susceptibility testing. The susceptibility of each strain to one or more of

184

the following antibiotics was determined before and after suspension test exposure to

185

a given formulation/microbicide following the British Society for Antimicrobial

186

Chemotherapy (BSAC) disk diffusion protocol (10): chloramphenicol (50 µg),

187

ampicillin (10 µg), ciprofloxacin (1 µg), ceftriaxone (30 µg), piperacillin (30 µg),

188

ceftazidime (30 µg), imipenem (10 µg), meropenem (15 µg), tobramycin (10 µg),

8


aztreonam (30 µg) (all from Oxoid, Baskingstoke, UK). These antibiotics were

190

selected due to their use as therapeutic agents in the treatment of infection with the

191

organisms chosen for this study. There are no available BSAC susceptibility

192

breakpoints for Burkholderia spp., so breakpoints for Pseudomonas spp. were used

193

instead in the case of strain UL2P (B. cepacia). (Approximate cost to evaluate

194

susceptibility of 1 strain to 6 antibiotics: < 2 €)

195 196

Phenotype stability testing. The stability of any alterations in antimicrobial

197

susceptibility observed after 5 min exposure of S. enterica strains SL1344 and

198

14028S to a range of low CHG and BZC concentrations was investigated via the 24

199

h subculture of surviving organisms through TSB +/- a low concentration of CHG or

200

BZC as described previously (3).

201 202

Data reproducibility. In order to determine the reproducibility of baseline and post-

203

exposure data obtained, the above experiments were performed on 3 separate

204

occasions (each using 3 biological replicates) over a 6 month period, resulting in data

205

values being a mean of 9 results.

206 207

Statistical analysis. A Students t-test was used to compare MIC, MBC and antibiotic

208

zone of inhibition sizes before and after microbicide exposure.

209 210

RESULTS

211

Three formulations and their corresponding microbicides were tested on three

212

separate occasions over a 6 month period in order to determine if exposure to a given

213

microbicidal product or microbicide resulted in an alteration in microbicide or

9


antibiotic susceptibility in test organisms. Data obtained on each occasion were

215

compared in order to determine the reproducibility of the MIC, MBC and antibiotic

216

susceptibility tests, and therefore validate the use of these tests as a high throughput

217

and low cost initial approach in the determination of the risk of resistance

218

development. The mean MIC and MBC for each test organism before and after

219

exposure to mouthwash, eye make-up remover or shampoo and their corresponding

220

microbicides (CHG, CHG, BZC) at the same concentration as that contained within

221

the product are presented in FIG.1. Exposure to one of three formulations or their

222

corresponding microbicides resulted in both increases and decreases in MIC and

223

MBC in individual strains. In the case of shampoo and eye make-up remover an

224

accurate MBC could not be determined as all 5 strains grew in the highest testable

225

concentration of the formulation. The greatest increases in MBC were observed in S.

226

enterica strain 14028S after exposure to 0.005 mg/mL CHG and mouthwash, and

227

0.015 mg/mL BZC, all of which were found to be significantly different from

228

baseline MBC values. However when considering the post-exposure MBC values

229

observed (0.08, 0.05 and 0.05 mg/mL respectively) it is clear that these values are

230

still below or equal to the concentrations of CHG and BZC present in the relevant

231

formulations when considered as a worst case scenario of product dilution by the

232

consumer. ‘Worst case’ dilution factors of 1 in 40 (mouthwash) and 1 in 100

233

(shampoo) were estimated based on product use, e.g. rinsing with water. This would

234

result in 0.05 mg/mL CHG in mouthwash and 0.05 mg/mL BZC in shampoo. An

235

MBC of 0.50 mg/mL for BZC is also of less concern as the primary function of BZC

236

in the shampoo is not as an antimicrobial, but as a surfactant. Very few of the

237

remaining observed changes in MIC or MBC were found to be statistically

10


significant (p≤0.05), nor did they approach the microbicide concentrations found in

239

the formulations tested after ‘worst case’ product dilution by the consumer.

240

An important factor in the validation of the use of MIC and MBC determination in

241

an initial assessment of the risk of resistance development was the reproducibility of

242

the data obtained. It is clear from FIG. 1 that both the baseline and post-exposure

243

mean MIC and MBC values were highly reproducible across the 3 separate

244

experiments, as indicated by the small standard deviations observed for each strain

245

and formulation/pure active. Our protocol is based on performing MIC/MBC in two

246

fold dilutions. Standard deviations were calculated based on the MIC or MBC

247

values, which means an increase or decrease in MIC or MBC by one fold dilution

248

will result in a large standard deviation. Error bars (representing SD) on the graphs

249

displayed in FIG. 1 may only indicate an increase or decrease of one doubling

250

dilution.

251 252

There was no clinical change in susceptibility to any of the antibiotics tested after 1

253

min exposure to all 3 formulations and their corresponding microbicides, in the case

254

of all 5 strains (according to BSAC susceptibility breakpoints for

255

Enterobacteriaceae/Pseudomonas spp. (10) (data not shown). In the case of some

256

strains and antibiotics, statistically significant changes in the zone of inhibition size

257

were observed. However these differences were often due to an increase in the mean

258

zone of inhibition size and therefore an increase in antibiotic susceptibility [e.g.

259

ciprofloxacin, chloramphenicol, ceftazidime in K. pneumoniae after exposure to

260

mouthwash (0.050 mg/mL CHG) or ceftazidime in P. aeruginosa after exposure to

261

shampoo (0.015 mg/mL BZC)]. A statistically significant reduction in the mean zone

262

of inhibition size for aztreonam was observed in P. aeruginosa after exposure to 11


0.005 mg/mL CHG, 0.015 mg/mL BZC and 1 mg/mL CHG. However P. aeruginosa

264

was already resistant to this antibiotic prior to microbicide exposure and therefore no

265

clinical susceptibility change was observed. It was not possible to clearly determine

266

if clinical changes in susceptibility were observed in B. cepacia, as there were no

267

available breakpoints provided in the BSAC protocol, and clinical susceptibility was

268

therefore based on Pseudomonas spp.

269

Carrying out this experiment on 3 separate occasions over a 6-month period also

270

allowed for an assessment of the reproducibility of the results obtained. The BSAC

271

method produces consistent and reproducible baseline and post-exposure data (data

272

not shown).

273 274

S. enterica strains SL1344 and 14028S were also exposed to a range of low

275

concentrations of CHG and BZC for 5 min before the antimicrobial susceptibility of

276

surviving organisms was determined. Tables one and two show the baseline and post

277

exposure values for SL1344 and 14028S respectively after 5 min exposure to a range

278

of low CHG and BZC concentrations.

279

In the case of both strains post-exposure MIC and MBC values for CHG and BZC

280

were all significantly different from baseline MIC and MBC values (p≤0.05). For

281

strain SL1344 the greatest increases in MIC and MBC were observed after 5 min

282

exposure to 0.004 mg/mL CHG and 0.004 mg/mL BZC (Table 1). For strain 14028S

283

exposure to 0.001 mg/mL CHG and 0.004 mg/mL BZC resulted in the greatest

284

increases in MIC and MBC in surviving organisms (Table 2). The data appear highly

285

reproducible across all 9 repeats in the case of both strains, as indicated by the low

286

standard deviation values, supporting our recommendation of the use of MIC and

287

MBC determination as an initial indicator of resistance development in bacteria. (As

12


discussed for FIG. 1, occasions where standard deviations appear larger are due to

289

the use of doubling dilutions of a given microbicide/formulation during MIC/MBC

290

testing). Susceptibility to a range of antibiotics was also determined for strains

291

SL1344 and 14028S before and after exposure to low CHG and BZC concentrations.

292

No alterations in antibiotic susceptibility were observed (data not shown).

293 294

The stability of the increases in MBC observed after 5 min exposure of SL1344 and

295

14028S to a range of low CHG and BZC concentrations was investigated via the 24

296

h subculture of surviving organisms through TSB +/- a low concentration of CHG or

297

BZC. Table 3 and 4 show the mean MBC values after 1, 5 and 10 subcultures of

298

surviving organisms through TSB +/- CHG or BZC for SL1344 and 14028S

299

respectively. The high MBC values observed after the initial 5 min exposure to CHG

300

or BZC were lost after 1 subculture in the absence of CHG or BZC. In the presence

301

of a low CHG or BZC concentration, MBC values also returned to baseline levels

302

after 10 subcultures. This was thought to be due to cumulative damage to the cell or

303

the fact that maintaining a high MBC was detrimental to cell survival. The instability

304

of the increased MBC values suggested a low risk of stable resistance development

305

to CHG or BZC in either S. enterica strain at the concentrations tested. The values

306

obtained from the phenotype stability tests were reproducible between repeats (as

307

indicated by the low standard deviation values in Tables 3 and 4) and the data

308

therefore supports our recommendation of the use this technique as part of a protocol

309

to predict microbicide resistance development.

310 311

DISCUSSION

13


The principle aim of this work is to design a protocol that can predict bacterial

313

microbicide resistance and antibiotic cross-resistance and give an indication of the

314

risk of resistance development. The purpose of this study was to validate the use of

315

MIC, MBC and antibiotic susceptibility determination before and after microbicide

316

exposure, and phenotype stability testing for use in the initial prediction of bacterial

317

microbicide resistance.

318

The use of existing standard protocols for MIC, MBC and antibiotic susceptibility

319

measurement (i.e. EN 1276, ISO 20776-1, BSAC disk diffusion method) helps to

320

avoid data variability which has been observed previously with MIC values obtained

321

using different methodologies. Schurmaans et al. (11) found that MIC values could

322

vary by a factor of up to eight if small alterations were made to the method used.

323

Phenotypic variability was avoided through the use of overnight broth cultures for

324

susceptibility testing, rather than selecting single colonies from an agar plate, which

325

has been demonstrated to result in phenotypic variability in Burkholderia cepacia

326

(12), illustrating the importance of consistent inoculum preparation when performing

327

susceptibility tests. In the work carried out here the inoculum was re-suspended in

328

diH20 instead of tryptone sodium chloride (TSC) buffer as TSC has been seen to

329

interfere with log reduction results due to carry over from the inoculum (unpublished

330

data). However the inoculum was used within 15 min of preparation in diH20 to

331

avoid subjecting bacterial cells to osmotic stress.

332

The MIC, MBC and antibiotic susceptibility values for mouthwash, shampoo, eye

333

make-up remover, CHG and BZC were found to be reproducible between separate

334

experiments at the concentrations tested in all 5 test strains, confirming the

335

appropriateness of using these standard protocols. We concluded that there is a very

336

low risk of resistance development to the formulations and corresponding pure

14


actives tested, even in the case of the elevated MICs and MBCs observed in strains

338

SL1344 and 14028S as these values were not stable in the absence or presence of

339

CHG or BZC.

340

The use of MIC and MBC in resistance prediction and making a comparison

341

between baseline and post-exposure susceptibility data is supported by our previous

342

work investigating the effect of cationic microbicide exposure on B. lata strain 383

343

(3). Our protocol allows the testing of any isolate of interest as data are always

344

compared for the individual isolate rather than general data for the given bacterial

345

species.

346 347

One of the criticisms of in vitro techniques used in microbicide resistance

348

measurement is that experimental parameters such as microbicide concentration,

349

exposure time, dilution on application and bioavailability are not reflective of in-use

350

conditions (1, 13). In our work we attempted to accurately reflect product use in

351

terms of exposure time and product concentration (i.e. any dilution of the product as

352

a result of its use). For the purpose of protocol development test concentrations used

353

were considerably lower than those found in the original formulations (i.e.

354

concentrations low enough to obtain surviving organisms), but should be kept

355

realistic when using the techniques recommended here to predict and assess the risk

356

of resistance development. Both formulations and the corresponding active

357

microbicides were tested during protocol development in order to validate the

358

different techniques used, but it must be emphasised that using such a protocol to

359

predict resistance to pure actives alone may be of less relevance than testing the

360

formulation as a whole, as multiple components of a formulation often contribute to

361

the overall microbicidal effect, or could prove antagonistic in the formulation.

15


Although better representative of microbicide use, long-term (≥ 6 months) studies

363

investigating the effect of exposure to commonly used household microbicides on

364

antimicrobial susceptibility, have failed to demonstrate resistance development in

365

isolated bacteria (14-17). These studies are also costly and do not allow for a prompt

366

response to regulatory bodies. This suggests that in light of new regulatory

367

expectations a compromise may be required, allowing the rapid generation of data

368

and preliminary assessment of risk, using in vitro techniques based on existing

369

standard methods whilst controlling parameters such as microbicide formulation,

370

contact time and concentration in order to bring realism to the evaluation. The

371

protocol proposed in this study aims to achieve this.

372 373

A further recommendation of Maillard et al. (1) and SCENIHR (2) in the event of

374

the observation of a reproducible change in microbicide susceptibility is the

375

execution of further tests to understand the nature of the change. This could include

376

molecular techniques to investigate changes to the transcriptome and proteome as a

377

result of microbicide exposure. Genotypic alterations as a result of microbicide

378

exposure and their potential as resistance markers have been investigated by

379

numerous groups (18-20), and molecular techniques such as PCR and microarray

380

technology have been successfully used to define microbicide resistance

381

mechanisms. Although useful, molecular techniques can be complex, costly and

382

time consuming and we therefore do not recommend them as a core part of this

383

predictive protocol. Taking this in to account, FIG. 2 shows the proposed protocol

384

steps in the form of a decision tree, as well as potential steps in the event of

385

observed, reproducible resistance. A stable increase in MIC or MBC or change in

386

antibiotic susceptibility could result in risk of resistance development. It must be

16


emphasised that the exact level of risk can only be determined through further

388

assessment. For example, a stable increase in MBC may not constitute a high level

389

of risk if this new MBC does not approach the concentration of a particular

390

microbicide intended for use (FIG. 2). Some microbicides have a long history of

391

use, and there is a large amount of literature studying their efficacy and any observed

392

bacterial resistance, e.g. chlorhexidine, triclosan, benzalkonium chloride. For these

393

microbicides there may be sufficient evidence available in the literature to support a

394

weight of evidence assessment of the risk of resistance development, before

395

considering the generation of new data on resistance (21, 22).

396 397

Our findings and proposed approach for assessment of risk can be applicable to the

398

wider use of microbicides in various settings where such compounds are applied.

399

This approach is preventative and aimed at being predictive, thereby ensuring that

400

microbicide-containing formulations are safe by design with regards to resistance

401

and cross-resistance risks, either by enabling omission of an ingredient identified by

402

the protocol as undesirable or by using the improved understanding of resistance and

403

cross-resistance mechanisms to design a formulation with an ingredient preventing

404

the expression of a microbicide-relevant resistance mechanism (e.g. efflux pump

405

inhibitors). Such a strategy has already been investigated and documented to

406

decrease bacterial resistance to antibiotics (23).

407 408

With regulatory bodies such as the US FDA and EU BPR requiring information on

409

the propensity of microbicidal products to select for resistant bacteria, it is

410

imperative that relevant, cost-effective, high throughput techniques are available in

411

order for product manufacturers to provide this information. As global harmonisation 17


of protocols used to measure changes in microbicide susceptibility is now considered

413

a key requirement in moving microbicidal research forward (1,2), we recommend,

414

and here demonstrate, the efficacy of a protocol that allows the prediction of

415

resistance development using simple, low cost and high throughput techniques.

416 417

Conflict of Interest

418

This project conducted by Cardiff University was sponsored by Unilever Safety &

419

Environmental Assurance Centre that provided a PhD studentship to L Knapp.

420 421

REFERENCES

422

1. Maillard J-Y, Bloomfield S, Coelho JR, Collier P, Cookson B., Fanning S,

423

Hill A, Hartemann P, McBain AJ, Oggioni M, Sattar S, Schweizer HP,

424

Threlfall J. 2013. Does microbicide use in consumer products promote

425

antimicrobial resistance? A critical review and recommendations for a cohesive

426

approach to risk assessment. Microb. Drug Resist. 19:344-354.

427

2. SCENIHR 2010. Scientific Committee on Emerging and Newly Identified Health

428

Risks: Research strategy to address the knowledge gaps on the antimicrobial

429

resistance effects of biocides. Available online at:

430

http://ec.europa.eu/health/scientific_committees/emerging/docs/scenihr_o_028.pd

431

f; accessed 24th November 2014

432

3. Knapp L, Rushton L, Stapleton H, Sass A, Stewart S, Amezquita A, McClure

433

P, Mahenthiralingam E, Maillard J-Y. 2013. The effect of cationic microbicide

18


exposure against Burkholderia cepacia complex (Bcc); the use of Burkholderia

435

lata strain 383 as a model bacterium. J. Appl. Microbiol. 115:1117-1126.

436

4. Christensen EG, Gram L, Kastbjerg VG. 2011. Sublethal triclosan exposure

437

decreases susceptibility to gentamicin and other aminoglycosides in Listeria

438

monocytogenes. Antimicrob. Agents Chemother. 55:4064-4071.

439

5. Escalada MG, Harwood JL, Maillard J-Y, Ochs D. 2005. Triclosan inhibition

440

of fatty acid synthesis and its effect on growth of Escherichia coli and

441

Pseudomonas aeruginosa. J. Antimicrob. Chemother. 55:879-882.

442

6. Mavri A, Smole MS. 2013. Development of antimicrobial resistance in

443

Campylobacter jejuni and Campylobacter coli adapted to biocides. Int. J. Food

444

Microbiol. 160:304-312

445

7. Thomas L, Maillard J-Y, Lambert RJ, Russell AD. 2000. Development of

446

resistance to chlorhexidine diacetate in Pseudomonas aeruginosa and the effect of

447

a "residual" concentration. J. Hosp. Infect. 46:297-303.

448

8. British Standards Institute. BS EN 1276: Chemical disinfectants and antiseptics:

449

Quantitative suspension test for the evaluation of bactericidal activity of chemical

450

disinfectants and antiseptics used in food, industrial, domestic and institutional

451

areas. Test method and requirements (phase 2, step 1). 20-9-2009.

452

9. BS EN ISO: 20776-1 (2006) International Organisation for Standardisation.

453

Clinical laboratory testing and in vitro diagnostic test systems - Susceptibility

454

testing of infectious agents and evaluation of performance of antimicrobial

455

susceptibility test devices - Part 1: Reference method for testing the in vitro

19


activity of antimicrobial agents against rapidly growing aerobic bacteria involved

457

in infectious disease.

458 459 460

10. Andrews JM. 2009. BSAC standardized disc susceptibility testing method (version 8). J. Antimicrob. Chemother. 64:454-489. 11. Schuurmans JM, Nuri Hayali AS, Koenders BB, ter Kuile BH. 2009.

461

Variations in MIC value caused by differences in experimental protocol. J.

462

Microbiol. Methods 79:44-47.

463

12. Larsen GY, Stull TL, Burns JL. 1993. Marked phenotypic variability in

464

Pseudomonas cepacia isolated from a patient with cystic fibrosis. J. Clin.

465

Microbiol. 31:788-792.

466

13. Harbarth S, Tuan SS, Horner C, Wilcox, MH. 2014. Is reduced susceptibility

467

to disinfectants and antiseptics a risk in healthcare settings? A point/counterpoint

468

review. J. Hosp. Infect..87: 194-202.

469

14. Aiello AE, Marshall B, Levy SB, Della-Latta P, Larson E. 2004. Relationship

470

between triclosan and susceptibilities of bacteria isolated from hands in the

471

community. Antimicrob. Agents Chemother. 48:2973-2979.

472

15. Aiello AE, Marshall B, Levy SB., Della-Latta P, Lin, SX, Larson E. 2005.

473

Antibacterial cleaning products and drug resistance. Emerg. Infect. Dis. 11:1565-

474

1570.

475

16. Cole EC, Addison RM, Rubino JR, Leese KE, Dulaney PD, Newell MS,

476

Wilkins J, Gaber DJ, Wineinger T, Criger DA. 2003. Investigation of

477

antibiotic and antibacterial agent cross-resistance in target bacteria from homes of

478

antibacterial product users and nonusers. J. Appl. Microbiol. 95:664-676.17. 20


17. Cole EC, Addison RM, Dulaney PD, Leese KE, Madanat HM, Guffey AM.

480

2011. Investigation of antibiotic and antibacterial susceptibility and resistance in

481

Staphylococcus from the skin of users and non-users of antibacterial wash

482

products in home environments. Int. J. Microbiol. Res. 3: 90-96

483

18. Bailey AM, Constantinidou C, Ivens A, Garvey MI, Webber MA, Coldham

484

N, Hobman JL, Wain J, Woodward MJ, Piddock LJ. 2009. Exposure of

485

Escherichia coli and Salmonella enterica serovar Typhimurium to triclosan

486

induces a species-specific response, including drug detoxification. J. Antimicrob.

487

Chemother. 64:973-985.

488

19. Furi L, Ciusa ML, Knight D, Di Lorenzo V, Tocci N, Cirasola D, Aragones

489

L, Coelho JR, Freitas AT, Marchi E, Moce L, Visa P, Northwood JB, Viti C,

490

Borghi E, Orefici G, Morrissey I, Oggioni MR. 2013. Evaluation of reduced

491

susceptibility to quaternary ammonium compounds and bisbiguanides in clinical

492

isolates and laboratory-generated mutants of Staphylococcus aureus. Antimicrob.

493

Agents Chemother. 57:3488-3497.

494

20. Yu BJ, Kim JA, Ju HM, Choi SK, Hwang SJ, Park S, Kim E, Pan JG. 2012.

495

Genome-wide enrichment screening reveals multiple targets and resistance genes

496

for triclosan in Escherichia coli. J. Microbiol. 50:785-791

497

21. SCENIHR 2012. Scientific Committee on Emerging and Newly Identified

498

Health Risks: Memorandum on the use of scientific literature for human health

499

risk assessment purposes - weighing of evidence and expression of uncertainty.

500

Available online at:

501

http://ec.europa.eu/health/scientific_committees/emerging/docs/scenihr_s_001.pd

502

f; accessed 24th November 2014 21


22. SCENIHR 2014. Scientific Committee for Emerging and Newly Identified

504

Health Risks: Nanosilver: safety, health and environmental effects and role in

505

antimicrobial resistance. Available online at:

506

http://ec.europa.eu/health/scientific_committees/emerging/docs/scenihr_o_039.pd

507

f; accessed 24th November 2014

508 509

23. Poole K. 2001. Overcoming antimicrobial resistance by targeting resistance mechanisms. J. Pharm. Pharmacol. 53: 283-294.

22

Applied and Environmental Microbiology 510TABLE 1: Mean baseline and post-exposure MIC and MBC values for strain SL1344 after 5 min exposure to a range of low CHG and BZC concentrations. N=9 511 Biocide concentration (mg/mL) ± SD MIC/MBC

Baseline

(mg/mL)

512

0.004

0.001

0.0005

0.0001

0.004

0.001

0.0001

513

CHG

CHG

CHG

CHG

BZC

BZC

BZC

514

CHG MIC

0.03 ± 0.03

0.80 ± 0.00

0.80 ± 0.00

0.40 ± 0.00

0.80 ± 0.00

0.50 ± 2.00

0.40 ± 0.00

515 0.80 ± 0.00 516

CHG MBC

0.10 ± 0.06

2.00 ± 0.90

2.00 ± 0.00

0.40 ± 0.00

1.00 ± 0.40

3.00 ± 0.00

2.00 ± 0.00

2.00 ± 1.00 518

517

519 BZC MIC

0.03 ± 0.00

2.00 ± 0.00

0.30 ± 0.20

0.10 ± 0.00

0.70 ± 1.00

3.00 ± 1.00

0.80 ± 0.00

0.70 ± 1.00520

BZC MBC

0.03 ± 0.03

2.00 ± 0.00

0.50 ± 0.20

2.00 ± 2.00

1.30 ± 2.00

8.00 ± 0.00

2.00 ± 0.00

3.00 ± 2.00

23

Applied and Environmental Microbiology 521 522TABLE 2: Mean baseline and post-exposure MIC and MBC values for strain 14028S after 5 min exposure to a range of low CHG and BZC concentrations. N=9

Biocide concentration (mg/mL) ± SD MIC/MBC

Baseline

(mg/mL ± SD)

0.005

0.001

0.015

0.004

CHG

CHG

BZC

BZC

CHG MIC

0.030 ± 0.03

0.10 ± 0.00

1.00 ± 0.00

0.40 ± 0.00

0.80 ± 0.00

CHG MBC

0.06 ± 0.03

1.00 ± 0.90

20.00 ± 0.00

50.00 ± 0.00

3.00 ± 0.00

BZC MIC

0.04 ± 0.03

0.80 ± 0.00

0.10 ± 0.00

0.80 ± 0.00

2.00 ± 0.00

BZC MBC

0.08 ± 0.02

1.00 ± 0.00

2.00 ± 0.60

1.00 ± 0.00

20.00 ± 0.90

24


TABLE 3: Mean baseline and post-exposure MBC values for strain SL1344 after 1, 5 and 10 subcultures in TSB +/- 0.004 mg/mL CHG or BZC.

524 SC = subculture

525

1 SC

*

= significantly different from baseline (p≤0.05)

Baseline

5 min CHG

MBC (mg/mL)

0.004

5 SC

0.10 ± 0.90

5.00 ± 0.00*

0.08 ± 0.00

0.09 ± 0.00

0.03 ± 0.00

1.50 ± 0.00*

0.04 ± 0.00

Baseline

5 min BZC

1 SC

MBC (mg/mL)

0.004

0.10 ± 0.90

5.00 ± 0.00*

0.20 ± 0.30

0.10 ± 0.00

0.03 ± 0.00

3.00* ± 0.00

0.06 ± 0.00

0.06 ± 0.00

10 SC

1 SC

5 SC

10 SC

(CHG)

(CHG)

(CHG)

0.06 ± 0.00

0.15 ± 0.40

0.10 ± 0.40

0.10 ± 0.00

0.06 ± 0.00

0.06 ± 0.00

0.19 ± 0.00*

0.50 ± 0.20*

0.06 ± 0.00

5 SC

10 SC

1 SC

5 SC

10 SC

(BZC)

(BZC)

(BZC)

0.10 ± 0.00

0.80 ± 0.40*

0.80 ± 0.40*

0.10 ± 0.00

0.06 ± 0.00

0.78 ± 0.00*

0.60 ± 0.20*

0.03 ± 0.00

CHG MBC (mg/mL ± SD) BZC MBC (mg/mL ± SD)

CHG MBC (mg/mL ± SD) BZC MBC (mg/mL ± SD) 526 527 528 529 530 25

Applied and Environmental Microbiology 531TABLE 4: Mean baseline and post-exposure MBC values for strain 14028S after 1, 5 and 10 subcultures in TSB +/- 0.004 mg/mL CHG or BZC. SC = subculture

532

*

= significantly different from baseline (p≤0.05)

533 Baseline

5 min CHG

MBC (mg/mL)

0.001

1 SC

5 SC

0.06 ± 0.03

5.00 ± 0.00*

0.01 ± 0.00

0.06 ± 0.00

0.08 ± 0.02

3.00 ± 0.00*

0.06 ± 0.00

Baseline

5 min BZC

1 SC

MBC (mg/mL)

0.004

0.06 ± 0.03

5.00 ± 0.00*

0.06 ± 0.00

0.05 ± 0.00

0.08 ± 0.02

3.00 ± 0.00*

0.07 ± 0.00

0.04 ± 0.00

10 SC

1 SC

5 SC

10 SC

(CHG)

(CHG)

(CHG)

0.09 ± 0.00

0.80 ± 0.40*

0.80 ± 0.40*

0.06 ± 0.00

0.07 ± 0.00

0.06 ± 0.00

0.19 ± 0.00*

0.20 ± 0.00*

0.06 ± 0.00

5 SC

10 SC

1 SC

5 SC

10 SC

(BZC)

(BZC)

(BZC)

0.06 ± 0.00

0.40 ± 0.20*

0.70 ± 0.70*

0.06 ± 0.00

0.06 ± 0.00

0.19 ± 0.00*

0.20 ± 0.00*

0.06 ± 0.00

CHG MBC (mg/mL ± SD) BZC MBC (mg/mL ± SD)

CHG MBC (mg/mL ± SD) BZC MBC (mg/mL ± SD) 534 535 536 537 538 26


27

Applied and Environmental Microbiology 540 541 542 543

FIG 1: MIC and MBC values for 5 test organisms re and after exposure to 3 formulations and their corresponding pure actives. N=9. Blue = baseline MIC. Red = postexposure MIC. Green = baseline MBC. Purple = post-exposure MBC. Error bars correspond to the SD. MIC and MBC were performed in two fold dilution (see text for detailed information). A) 0.005 mg/ml CHG; B) mouthwash (0.005 mg/mL CHG); C) 1 mg/mL CHG; D) Eye-maker remover (neat: 1 mg/mL CHG); E) 0.015 mg/mL BZC; F) Shampoo (0.015

544 545

28

Applied and Environmental Microbiology 546Figure 3: Proposed protocol for use in the prediction of bacterial microbicide resistance. Grey boxes are examples of further work that could be carried out to investigate 547mechanisms behind changes in antimicrobial susceptibility. 548 549 No

550

MIC/MBC/antibiotic susceptibility Is there an increase in MIC/MBC? Is there a change in antibiotic susceptibility? After exposure to a product under realistic conditions1

551 552

Yes

Low Risk

553

Decision point: • Evaluate increased MIC/MBC against realistic, in-use concentrations • Compare decreased antibiotic susceptibility against clinical breakpoint

554 No

555

Phenotype stability testing Are the observed changes stable?

556 Yes

557 558

Further investigate risk2

559 560 561 562

Mechanisms of resistance?

563

Microarray Identification of potential marker genes

Efflux assays Does efflux activity increase?

Membrane protein expression Change in outer membrane proteins?

Real time PCR Confirmation of microarray changes Identification of marker gene

29

Applied and Environmental Microbiology 564Footnotes for figure 3 5651 Realistic conditions refers to those under which the product will be used. Factors such as concentration, contact time and product formulation should be considered in 566order to represent product use as accurately as possible. 5672 If reproducible and phenotypically stable changes in antimicrobial susceptibility are observed after exposure to a particular product under realistic, in-use conditions, 568further investigation into the risk can be carried out. This may involve the elucidation of possible mechanisms behind susceptibility changes such as the examples shown in 569the grey boxes in figure 3, leading to better understanding of the level of risk. This investigation could be extended beyond the examples given in figure 3, and could 570include the exploration of additional resistance markers and the use of additional techniques. 571

30